In-Depth Review: xFigura
What It Is
A shared canvas for architectural concept work that puts more than 90 third-party image, video, and 3D models behind one subscription, with an agentic chat that reads what is on the canvas and generates new views from it.
Who It's For
Design teams doing concept and visualization work, from solo practice to global firms. The company names OMA, Zaha Hadid Architects, KPF, BIG, Snohetta, and Herzog and de Meuron as users.
Why It Matters
The models behind every AI rendering tool change every few months, so a studio that buys a single-model product is tied to whichever engine that product wrapped. xFigura sells access to 50 image models, 15 video models, and 10 3D models under one subscription and one dashboard, which makes the choice of engine a per-job decision rather than a procurement one. The canvas matters for a second reason. Concept work is comparative, and placing variations next to each other with comments on them matches how a design review runs better than a queue of one-off renders does.
Problem Solved
A studio exploring a direction ends up with images in one tool, a video in another, a massing study in a third, and the review happening over email. xFigura keeps the whole exploration on one board. You bring a mood image, a sketch, or a 3D view in from the tools you already use, generate massing and photoreal views from it, annotate and comment directly on images, video, and 3D, and keep the variations beside each other. The agentic chat reads an attached render, extracts the palette, glazing lines, and depth, then plans and composes additional views that hold the same light and materials.
Audience Fit
Primary Users:
- Architect
- Designer
- Visualization Artist
- Student
Ideal Fit:
- Concept and schematic work where many directions get compared
- Studios that would otherwise buy several separate AI rendering subscriptions
- Teams that review visuals together rather than one person generating alone
- Firms wanting one dashboard with usage tracking across many designers
Key Differentiator
More than 90 image, video, and 3D models on one collaborative canvas
Workflow Fit
Project Phases Covered:
- Feasibility
- Concept
- Schematic
Where It Plugs In:
Sits before and beside modeling. It imports images, views, and 3D models from the tools a studio already runs, and everything it returns is a visual rather than measured geometry, so the documentation set stays where it is.
User Experience
Speed: Browser-based with nothing to install. Generation time depends on which of the 90 models a job uses
Interface: An infinite canvas holding images, video, and 3D side by side, with commenting and annotation on each, plus an agentic chat panel that works on what is attached
Learning Curve: Generating a first image takes minutes, and the harder skill is knowing which of the 90 models suits a given job
Collaboration: Shared boards, collaborator access, public board visitors on Professional, a shared credit pool and admin dashboard on Team, and SSO with firm-wide usage tracking on Enterprise
Integrations: Imports images, views, and 3D models from existing design tools. Enterprise adds SSO, SAML, and customized API access
Bonus Insight
xFigura describes its commercial licence two different ways. The pricing table labels the usage licence on Starter and Professional as Personal and reserves Royalty-free for Enterprise, while the FAQ on the same page says generations can be used in client projects on any paid plan and that the only commercial-use restriction is on the Free plan. Those statements conflict, and the answer decides whether a $16 plan covers billable work. The data terms contain another conflict. The homepage says no model retraining, ever, and the Terms of Service say xFigura receives a licence to use customer content to maintain, improve, and train its products, with enterprise terms restricting training on client visual content. Credits meter every generation against one pool and video and 3D cost more than an image, so the 400 credits on Starter produce 400 images, 130 videos, or 100 3D models. The site describes its security as SOC 2 aligned rather than certified and names no audit type, auditor, or report. The company publishes a named quote from Patrik Schumacher at Zaha Hadid Architects and case studies with ZHA, KPF, and BIG. Its performance figures have no published methodology: 500,000 generations a month, $1 million saved on visualization at one practice in six months, and 75 percent faster concept iteration. The product returns visuals rather than measured geometry, so project documentation cannot be produced from anything it generates.
Need help implementing xFigura?
Our advisory team has hands-on experience across 520+ AEC tools. Book a free AI strategy session to get specific recommendations for your firm.